Tlie Pre-Glacial Valleys of Arran and Snoivdon. 155 



Willis (1893, pi. Ixxv). Such tension clefts would be narrow, but 

 would be enlarged by the streams that would flow along them. 



Broader valleys would be formed v/here tension is relieved by 

 a multiple instead of by a simple fracture. This multiple fracture 

 might consist of parallel fractures connected by cross fractures, or 

 of anastomosing branching fractures. Such multi|)le fractures would 

 produce a belt of shattered rock and thus form one variety of the 

 shatter-belts of Professor Marr. His definition (1906, p. 77 ; re- 

 affirmed 1916, p. 71) attributed shatter-belts to oscillatory faults, 

 which cause no final displacement as the rocks on both sides are 

 restored to their original positions by the backward and forward 

 movement. He describes these belts as many yards in width and 

 as composed of rock " broken into a rubbly mass of angular frag- 

 ments of varying size in a matrix of finer crushed material ". Such 

 belts of shattered material may also be formed by bands of rock 

 being torn to pieces by tension cracks ; and some shatter-belts of the 

 Lake District would appear to be more easily explained as multiple 

 anastomosing tension clefts than as faults ; and the broader valleys 

 of Arran and the major valleys of the rectilinear system around 

 Snowdon may occur along shatter-belts due to tension during the 

 Pliocene uplift. Shatter-belts probably often occur where they cannot 

 be seen, for they would be seldom exposed on the floors of the valleys, 

 and quickly obscured on the sides ; and in the absence of clear ex- 

 posures, shatter-belts formed as crush breccias are indistinguishable 

 from those due to multiple tension fractures. The term shatter-belt 

 may be conveniently used for both kinds. 



4. Amount of Glacial Excavation in the Valleys. 



With the increasing severity of the climate the country was 

 occupied by glaciers, and they found an immature topography on 

 which even limited glacial erosion would have produced conspicuous 

 results. The iqe flowing down the valleys must have pressed against 

 the spurs, cut them back, and thus rendered the north and south 

 valleys trough-shaped. There seems no convincing evidence that 

 the ice materially deepened the valleys. Some of the alluvial plains 

 may cover buried channels formed at the same time as' those on the 

 mainland ; but none of the freshwater lochs of Arran have floors 

 lower than sea-level. No known Arran valley has been excavated 

 below the base-level of the pre-Glacial rivers. Accordingly there is 

 no need to assume the glacial deepening of the valleys. 



The amount of hard rock removed by ice in widening the valleys 

 was not necessarily considerable. The ice doubtless wore away the 

 spurs, some of which are now represented only by blunt gables on 

 the hill-sides, such as Torr Breac at the bend of Glen Rosa. The 

 smooth spurless slopes at first give the impression of extensive ice 

 erosion. But if, as seems most probable, the three north and south 

 valleys were formed along tension clefts made by the uplift, their 

 original course would have been straight and the spurs on their 



